Compact and Robust Deep Learning Architecture for Fluorescence Lifetime Imaging and FPGA Implementation

TL;DR

This paper introduces a novel, efficient deep learning architecture for fluorescence lifetime imaging that reduces computational complexity and is suitable for FPGA implementation, maintaining high accuracy with significant data compression.

Contribution

The paper proposes a multiplication-free 1-D deep learning network (FLAN) with a log-scale merging technique for FLIM, optimized for FPGA deployment, offering improved efficiency and comparable accuracy.

Findings

FLAN and FLAN+LS achieve high accuracy in synthetic and real data.

FLAN+LS provides significant data compression ratios.

FPGA implementation enhances computational efficiency.

Abstract

This paper reported a bespoke adder-based deep learning network for time-domain fluorescence lifetime imaging (FLIM). By leveraging the l1-norm extraction method, we propose a 1-D Fluorescence Lifetime AdderNet (FLAN) without multiplication-based convolutions to reduce the computational complexity. Further, we compressed fluorescence decays in temporal dimension using a log-scale merging technique to discard redundant temporal information derived as log-scaling FLAN (FLAN+LS). FLAN+LS achieves 0.11 and 0.23 compression ratios compared with FLAN and a conventional 1-D convolutional neural network (1-D CNN) while maintaining high accuracy in retrieving lifetimes. We extensively evaluated FLAN and FLAN+LS using synthetic and real data. A traditional fitting method and other non-fitting, high-accuracy algorithms were compared with our networks for synthetic data. Our networks attained a minor reconstruction error in different photon-count scenarios. For real data, we used fluorescent beads' data acquired by a confocal microscope to validate the effectiveness of real fluorophores, and our networks can differentiate beads with different lifetimes. Additionally, we implemented the network architecture on a field-programmable gate array (FPGA) with a post-quantization technique to shorten the bit-width, thereby improving computing efficiency. FLAN+LS on hardware achieves the highest computing efficiency compared to 1-D CNN and FLAN. We also discussed the applicability of our network and hardware architecture for other time-resolved biomedical applications using photon-efficient, time-resolved sensors.